Summary Under the proposed Cloud and AI Development Act (CADA), researchers can access critical AI computing resources through a targeted mechanism for "frontier AI priority projects." As proposed in Articles 8 and 9, the Union and Member States would establish a matching framework where the EU contributes compute time from its share of European High-Performance Computing (EuroHPC) capacity to projects that meet strict strategic criteria. To qualify, a project must be a pioneering effort in frontier AI, undertaken by a European digital infrastructure consortium (EDIC) or eligible entity involving at least three Member States, with those states pooling resources. This support is operationalized alongside a network of "Experience and Acceleration Centres for AI," which serve as regional gateways for researchers to access testing environments, expertise, and the broader EU innovation ecosystem.

Detail

The proposed Cloud and AI Development Act (CADA), COM(2026) 502 final, addresses a critical bottleneck identified in the Commission's explanatory memorandum: the "unprecedented and growing demand for computational capabilities" driven by AI proliferation. The proposal recognizes that computing infrastructures have evolved from mere technical assets into "strategic resources critical to the Union's economic security, sovereignty, resilience, and competitiveness." To mitigate the risks of dependence on third-country providers and to foster a competitive European AI ecosystem, CADA introduces a specific legal framework in Articles 8 and 9 to allocate high-performance computing resources to strategic research and development.

The Frontier AI Priority Project Mechanism (Article 8)

Article 8 establishes the criteria for the Commission to formally recognize projects as "frontier AI priority projects." This designation is not automatic; it is granted through "open calls for expressions of interest" and is strictly tied to Grand Challenge 3 set out in Annex I of the regulation. Grand Challenge 3 focuses on "Developing the next generation of multimodal frontier AI models and systems and pioneering novel capabilities," including architectural designs that push boundaries in advanced reasoning, cross-modal understanding, and agentic capabilities.

To be recognized as a frontier AI priority project, a proposal must fulfill three cumulative criteria as set out in Article 8:

  1. Pioneering Focus: The project must be a "pioneering project, focused on the support and scaling-up of frontier AI technologies." This excludes incremental improvements or standard applications, targeting instead the development of strategic assets that maintain a competitive edge in the global digital economy.
  2. Entity and Multi-State Participation: The project must be undertaken by a "European digital infrastructure consortium established pursuant Decision (EU) 2022/2481 or another legal entity eligible for funding under Union law." Crucially, it must "involve the participation of at least three Member States." This requirement ensures that frontier AI development is a collective Union effort, preventing fragmentation and leveraging cross-border expertise.
  3. Resource Pooling: The participating Member States must "pool computing time and other relevant resources to support the implementation of the designated project." This criterion mandates that national contributions are not merely symbolic but constitute a tangible commitment of compute capacity.

This mechanism is designed to address the "technical complexity and capital-intensive nature" of frontier AI. By requiring multi-state participation and resource pooling, CADA aims to create a unified front for developing strategic AI assets, particularly in sectors like cybersecurity where the explanatory memorandum notes the need to "reduce current dependencies on third-country technologies."

Computing Support and the EuroHPC Matching Framework (Article 9)

Article 9 translates the designation of priority projects into concrete computing support. It imposes obligations on both the Union and Member States to ensure that "sufficient AI computing resources from their compute capacities are allocated to support the development of frontier AI priority projects that fulfil the criteria set out in Article 8, within the limits of available capacity."

The core of this support is a matching mechanism:

  • Union Matching: "The Union shall at least match the AI computing resources contributed by Member States to frontier AI priority projects." This matching is conditional, however: it applies "to the extent that sufficient AI computing capacity is available within the Union's share of European high performance computing access time."
  • EuroHPC Integration: The proposal explicitly links this support to the EuroHPC Joint Undertaking (EuroHPC JU). The text states that "The EuroHPC JU access policy should be accommodated to reflect the allocation of such computing resources in an efficient, transparent and timely manner." This accommodation is "without prejudice to the continuity of ongoing operations and the rights of projects already benefiting from allocated EuroHPC AI computing resources." This ensures that CADA does not disrupt existing supercomputing commitments but rather carves out a strategic priority lane within the existing EuroHPC framework.

Beyond frontier AI, Article 9(3) extends a broader, albeit less binding, obligation: "The Union and the Member States shall endeavour to provide sufficient computing resource for AI industrial innovation, physical AI and public sector AI projects." This indicates that while frontier projects receive the "matching" guarantee, other strategic areas (such as physical AI for robotics or industrial AI for manufacturing) are also prioritized for resource allocation, subject to availability.

The Role of Experience and Acceleration Centres for AI

While Articles 8 and 9 govern the allocation of raw compute power at the Union level, Article 5 establishes the operational infrastructure that connects researchers to these resources. The proposal mandates that "Each Member State shall establish Experience and Acceleration Centres for AI ('Centres for AI')." These centres build on the existing network of "European digital innovation hubs" (EDIHs) and serve as regional and local accelerators.

For researchers, these centres are critical entry points. Their tasks include:

  • Access and Transformation: "Helping organisations accelerate their digital transformation through access to and use of AI technologies, including by connecting organisations with European providers of cloud and AI technologies."
  • Testing and Fine-Tuning: "Leveraging relevant infrastructure to accelerate the development and fine-tuning of AI models and systems." This suggests that the Centres will provide the necessary testing environments and pilot lines that complement the high-performance compute allocated under Article 9.
  • Skills and Expertise: "Ensuring or providing access to relevant upskilling and reskilling schemes, in close collaboration with the AI Skills Academy," and "facilitating the transfer of expertise across regions."

The Centres for AI are thus the "last mile" of the CADA compute strategy, ensuring that the high-level allocation of EuroHPC time under Article 9 is accessible to researchers who may lack the administrative capacity to navigate Union-level calls directly.

Links to the Broader Ecosystem and Sovereignty

The proposal positions CADA as a complement to existing EU initiatives, particularly the EuroHPC JU and the Chips Act. The explanatory memorandum notes that the Cloud and AI Leadership Initiatives (Title II) should "foster the co-design and cross-optimisation of hardware and software development and the integration of AI computing infrastructures" alongside the Chips for Europe Initiative 2.0.

Furthermore, the proposal emphasizes the importance of open source and European cloud stacks. The Cloud and AI Leadership Initiatives support the development of "open cloud computing stacks" and "AI-optimised servers and baseline software based on processors, accelerators and quantum accelerators designed and manufactured in the Union." This ensures that the compute resources provided under Article 9 are not only abundant but also aligned with the Union's goals of "technological autonomy" and "supply chain resilience."

The sovereignty aspect is also implicit in the criteria for Article 8. By requiring projects to be undertaken by entities established under Union law and involving multiple Member States, the proposal ensures that the resulting AI models and systems remain under Union control, reducing the risk of "extraterritorial effects of legislation adopted by third countries" mentioned in the recitals.

What this means for you

For researchers, CTOs, and academic institutions evaluating the practical impact of CADA, the proposal offers a structured, albeit competitive, pathway to access scarce and expensive AI computing resources.

  1. Form Multi-State Consortia: Individual researchers or single-entity projects are unlikely to qualify for "frontier AI priority project" status. Article 8 explicitly requires the participation of "at least three Member States." You must proactively build partnerships across borders, potentially forming or joining a European digital infrastructure consortium (EDIC).
  2. Align with Grand Challenges: Your research proposal must align with Grand Challenge 3 (Frontier AI) or other strategic challenges in Annex I (e.g., Physical AI, Industrial AI). Projects focusing on "advanced reasoning, cross-modal understanding and agentic capabilities" are the primary targets. General AI research may still access EuroHPC resources, but the specific "matching" mechanism in Article 9 is reserved for these strategic priorities.
  3. Engage with Local Centres Early: Utilize the Experience and Acceleration Centres for AI in your region. These centres are tasked with "connecting organisations with European providers" and "facilitating the transfer of expertise." They can help you navigate the application process, identify potential partners in other Member States, and prepare the necessary technical documentation for the open calls.
  4. Prepare for EuroHPC Integration: If your project qualifies, you will access compute through the EuroHPC JU framework. Familiarize yourself with the EuroHPC access policies, as Article 9 mandates that these policies be "accommodated" to reflect CADA allocations. Be prepared to demonstrate how your project contributes to the Union's technological sovereignty and how your Member State is willing to "pool computing time" to trigger the Union's matching contribution.
  5. Leverage the Matching Mechanism: The "at least match" provision in Article 9 is a powerful lever. If your consortium can secure a commitment from three Member States to pool resources, the Union is obligated to match that contribution (subject to capacity). This effectively doubles the available compute for your project, significantly amplifying your research capabilities.

Common misconceptions

"Any AI research project can apply for compute support." No. Support is targeted at "frontier AI priority projects" that meet specific, cumulative criteria under Article 8, including multi-state participation and alignment with Grand Challenge 3. General AI research may still access EuroHPC resources through existing channels, but the specific matching and prioritization mechanisms in Articles 8 and 9 are reserved for strategic, pioneering projects.

"CADA replaces the EuroHPC Joint Undertaking." No. CADA complements and leverages the EuroHPC JU. It does not create a new supercomputing infrastructure but rather establishes a policy framework for allocating existing EuroHPC capacity to priority AI projects. The proposal explicitly states that the EuroHPC JU access policy should be "accommodated" to reflect CADA allocations, ensuring continuity for existing projects.

"Access is guaranteed for all EU-based researchers." No. Access is competitive and conditional. Projects must be selected through "open calls for expressions of interest" and must fulfill the strict criteria for frontier AI priority projects. Furthermore, the Union's matching contribution is "within the limits of available capacity," meaning that if EuroHPC capacity is fully utilized, the matching may be constrained.

"SMEs are excluded from this support." While the "frontier AI priority project" status requires significant scale and multi-state participation, SMEs can participate as partners in these consortia. Additionally, the Centres for AI are explicitly tasked with supporting SMEs and SMCs in their digital transformation and access to AI technologies, providing a separate pathway for smaller entities to benefit from the ecosystem.

Official sources

Related

This is general information about a draft EU regulation, not legal advice.